[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Full Bandwidth Color Difference



Martin wrote:

> full bandwitdth RGB is certainly better than YUV no matter what your
> working on.


I'm sorry... are you saying that (to use a realistic example)  10 bit 
4:4:4
RGB is better than 10 bit 4:4:4 YCbCr ?

Yes because 10 bit 4:4:4: YCbCr has a problem (granted very small) in
getting the chroma/luminace ratio right. The Cb and Cr signals are half of
the Y signal and must end up being equal to work properly in the display
device. If the gain of the CbCr is not exactly on then saturation is not
right. Again this problem is usually seen on varios display devices.
Clients however find this an asset being able to super saturated or
undersaturate colors or going black & white from color images. With RGB
that creative factor of exagerating saturation is not a convenient option.
The color you get on a properly setup machine is as it was photographed. In
those telecine machines that use 10 bit 4:2:2 another factor enters in
"constant luminace". Most ignore this but it is there.

Also, if you are working on a Quantel box, where YCbCr is the native format
you are ahead of the game because you avoid color-space conversions during
in/out.  In this case RGB wouldn't be better, right?

Yes, I agree that you don't have to make conversions that have tolerances
and can lead to chroma/luminace inequities.

Other than color-space conversion issues what are we talking about here?

At one point my rule of how to buy guitars might be applicable here:  If
you
can't hear a difference don't pay for the difference.

Unfortunately what we can see is not good enough for post production. The
eye sees fine detail in Black & White so the idea of half the sampling for
color is what TV is about. Be it HDTV or SDTV. but if you are doing FX the
FX device can, especially in a color diference situation, see the
difference.

At what point is all this technical stuff nothing less than digit-head
nitpicking (I am one, so I can say this).  A finished product is intended
to
be consumed by humans, not machines.  If human perception deems it adequate
or better than what was before... well, how much more do we argue about
loosing a bit here or there.  The MPEG, JPEG, etc. folks figured all this
stuff out a while ago.

Correct, MPEG, JPEG, etc are distribution formats. They were not intended
to be used for production or post production. MPEG or JPEG were to look at
not manipulate.

If we wanted to be as close to techical perfection as possible --the goal
being to capture as much of the information contained in the negative for
further processing--  we should all be using Cineon (or Imagica, or
whatever) scanners at 4K  and 14 bits linear.  If we are talking about
getting shows, MOW's, promos, commercials, etc. done in a flowing real-time
(or near) environment any of the currently available technologies are more
than adequate for the task at hand (In my humble opinion).

And in my humble opinion also!


Jim Mendrala
Real Image Technology, Inc.
(805) 294-1049
mailto:J_Mendrala at compuserve.com

---
Thanks to Peter Stansfield for supporting the TIG in 1998..
No product marketing allowed on the main TIG.  Contact rob at alegria.com
946 subscribers in 36 countries on Wed Mar 25 09:30:13 PST 1998
complete information on the TIG website http://www.alegria.com/tig3/